Every professional investor knows that they must read the financial statements of a company to get a deep understanding of a company, its operations and its financials. But who among us has ever analysed the financial statements of an equity fund? Maybe we should because there is some useful information in there to predict future fund performance.
James Li from the Wharton Business School examined if the information in the financial statements of open-end and closed-end funds in the US provides insights above and beyond what one can get from the usual data like expense ratios, past performance, and portfolio turnover. To do this, he looked at the financial statements of 3.156 open-ended domestic US equity funds (all actively managed) and 56 closed-end funds (also all actively managed). In total these are some 23,000 financial statements between 1995 and 2018.
From these financial statements, he created ten accounting variables, such as the dividend and interest income earned by a fund, the cash holdings, the income from securities lending and any realised and unrealised gains and losses. He then combined them with traditional metrics used to evaluate mutual funds like total expense ratios, portfolio turnover, active share, fund size, etc.
He found that several accounting measures help an investor better predict future performance by the fund:
Dividend income: Funds with higher dividend income tend to outperform funds with lower dividend or interest income in the future.
The same holds true for funds with higher interest income.
Unrealised gains: Funds with higher unrealised gains are likely to underperform funds with lower unrealised gains in the future. Paper gains tend to disappear quickly as do paper losses, which is why funds with higher unrealised losses tend to outperform funds with lower unrealised losses.
While not predictive of future fund performance, cash holdings and expense reimbursements are predictive of the relative performance between funds. Funds with higher cash levels tend to perform better than funds with lower cash holdings (which may surprise readers) and funds that temporarily waive some of their fees for investors tend to outperform other funds that do not (which is just another way of saying that funds with lower fees outperform funds with higher fees).
Using these accounting measures, Li could create a machine learning algorithm that increased the alpha of his fund selection from 1.4% to 2.4% per year vs. a model that only uses standard fund metrics. That’s not bad and it seems to be something that investors are completely oblivious about. He tested if investors in mutual funds redirect investments in reaction to these funds publishing their financial statements and could not find and evidence that this was happening. Similarly, closed-end fund discounts to NAV did not change after the funds published their financial statements, indicating that closed end fund prices also do not react to the information given in their financial statements.
Hence, if you are a professional fund selector or just an investor looking for high-performing actively managed funds, here is information that is easily available to everyone, yet completely ignored by the market. Doing the extra work can give you a real edge in this space.